Predicting construction project durations in Europe using regression analysis methodology

The United States Corps of Engineers (USACE) engages the expertise and services of external and internal stakeholders to complete construction projects safely in high-quality form without overrunning schedules and budgets. However, completing construction projects on time remains challenging in today’s rapidly changing business environment. This research aims to overcome the challenge of exceeding construction project completion durations in Europe by developing quantitative-based predictive models instead of solely relying on experiences.This study utilizes the Multiple Linear Regression Analysis statistical methodology to analyze 291 completed construction projects in the USACE Europe District. A total of 16 independent variables that could influence completion durations are identified in the Resident Management System (RMS) database. The research identifies significant predictors for each project scale using stepwise selection. Four significant variables emerge for medium and large-scale projects’ predictive model: Original Value w/ Options, No. of Transmittals, No. of Modifications, and Changes over $250K, achieving an R-squared value of 93.96%. The small-scale project model employs five predictors: Original Value w/ Options, No. of RFIs, No. of Transmittals, Changes over $50K, and Changes over $150K, demonstrating an R-squared value of 95.48%. Both models undergo rigorous validation through 5-fold cross-validation, confirming their robust predictive capabilities with cross-validated R-squared values of 91.44% and 91.93%. These models provide USACE project and program managers with reliable tools for predicting construction project completion durations during early project phases, enabling better resource allocation and risk management. The dual-model approach ensures higher prediction accuracy by accounting for the distinct characteristics of different project scales, ultimately supporting the District's increased mission readiness and stronger stakeholder relationships.